DESCRIPTION Abstract: Despite decades of effort, no current vaccine elicits neutralizing antibodies at concentrations blocking HIV infection. In addition to structural features of HIV's envelope spike that facilitate antibody evasion, we propose that the low density and limited lateral mobility of HIV spikes impedes bivalent binding by antibodies. The resulting predominantly monovalent binding minimizes avidity and thereby high affinity binding and potent neutralization, thus expanding the range of HIV mutations permitting antibody evasion. The HIV spike trimer geometry does not allow intra-spike cross-linking by naturally-occurring bivalent antibodies, but we can construct proteins capable of high-avidity trivalent binding to a spike for gene therapy and/or passive immunization. We will design, express, and test trimeric intra-spike cross-linking reagents that bind to two or three non-overlapping sites per spike monomer (6-9 sites per trimer). Choosing HIV-binding proteins and how to link them will be done combinatorially starting with a library of 15-30 HIV spike-binding proteins coupled to double-stranded DNA identifying tags. These will be linked using variable-length DNA to form bispecific reagents separated by different distances. Pooled bispecific reagents will separated by affinity chromatography to isolate the tightest binding pairs, which will be PCR amplified/sequenced to determine the two protein components and the linker length. Upon identifying the optimal length for the linker, the DNA is replaced with a comparable-length protein linker, and the two-binding-protein reagent is linked to a third binding protein from the HIV-binding library. Optimal two- or three-component HIV binding proteins will be trimerized by attaching a trimerization motif first via DNA linkers to determine the optimal length, then using protein-based linkers. The trimeric reagent's intra-spike cross-linking would reduce the concentration required for sterilizing immunity, making HIV's low spike